From the NVIDIA edition of October 29, 2026
NVIDIA and EPFL: cuBLAS FP32 emulation on Tensor Cores speeds machine-learning interatomic potentials by up to 1.7x on Blackwell and 1.5x on A100 with one environment variable
NVIDIA with EPFL (Michele Ceriotti) · Scientific computing (vendor technical blog) · Global (Switzerland and United States contributors)
- Publisher
- NVIDIA Technical Blog
- Original publication
- 2026-10-29
- Source retrieved
- 2026-10-30
What happened
NVIDIA's Oct 29 technical post, co-authored with EPFL's Michele Ceriotti, reports that cuBLAS FP32 emulation can run single-precision matrix math on Tensor Cores without code changes, enabled by setting CUBLAS_EMULATE_SINGLE_PRECISION=1. On Blackwell, BF16x9 emulation sped large GEMMs in the PET model by 2.2-2.5x for a 17,280-atom system and the full molecular dynamics step by up to 1.7x. On A100, TF32x3 sped GEMMs by 1.7-1.9x and the MD step by up to 1.5x. The authors say end-to-end gains depend on system size and GEMM share, and TF32x3 should be validated against accuracy requirements.
Why it matters
University and national-lab-affiliated materials science groups on existing A100 and Blackwell clusters may get faster simulations without new hardware, extending the value of public research computing investments.
Evidence and measured results
Vendor technical blog with co-author benchmarks on specific models and systems; no independent replication.
Limitations and uncertainty
Vendor-authored with benchmarks on a single model family; gains vary with system size; TF32x3 trades some accuracy and must be validated; combined gains with other optimizations were not measured.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-30; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Research groups need faster simulations but capacity budgets are limited.
- Stakeholders
- Research computing director
- Materials science faculty
- Discovery questions
- Which user codes are dominated by FP32 GEMMs?
- What GPU generations are in the cluster?
- Value hypothesis
- Library-level emulation may increase throughput on existing GPUs.
- Potential engagement
- Benchmarking support for user codes with emulation enabled.
- Unsupported claims
- Gains for models beyond those tested
- Accuracy equivalence without validation
Pre-sales engineering
Role takeaway
- Fit
- Clusters running FP32-heavy scientific ML codes on Ampere or Blackwell GPUs.
- Architecture and integration
- Environment-variable setting in cuBLAS, delivered via environment modules.
- Prerequisites
- Supported cuBLAS/driver versions
- FP32 baseline for comparison
- Constraints
- Gains depend on system size and GEMM share
- TF32x3 accuracy trade-off
- Security
- No direct security implications.
- Proof of value
- Benchmark speedup and accuracy on representative codes
- Validate trajectory stability for TF32x3
Delivery
Role takeaway
- Work
- Benchmark emulation with research groups and publish guidance.
- Dependencies
- Research groups
- Cluster software stack
- Ownership
- Research computing user support.
- Skills and adoption
- Guidance on numerical validation for researchers.
- Governance checkpoints
- Benchmarks reviewed with researchers
- Guidance published
- Acceptance criteria
- Proposed: measured speedup with accuracy within user-defined tolerance
- Proposed: documented settings in job scripts
- Risks
- Silent accuracy loss
- Version incompatibilities
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Precision emulation is a library-level setting, so cluster teams can offer it through environment modules without changing user code.
Governance
Who approves, reviews and stays accountable for outcomes?
Research groups should document numerical settings for reproducibility.
Security and privacy
What data, permissions and controls need testing?
No direct security implications identified.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Researchers need guidance on when emulation is numerically appropriate.
Procurement
What should contracts, pricing and exit terms secure?
May improve throughput of already purchased GPUs, which can inform capacity planning.
Operating model
Which teams own the service once it runs?
Cluster support teams can publish tested settings and validation guidance for common codes.
Publication history
- 2026-10-29NVIDIA · Issue 547 resources
Stable resource ID: nvidia-devblog-tensor-core-emulation-ml-interatomic-potentials-epfl-20261029